Calculate average of multiple matrices by reading an m.file

I'm having difficulty in calculating the mean of multiple matrices. I know how to calculate the mean, I have tried to use mean(match,3) in 3 dimension , but it does not give me the mean of all my matrices by columns If I calculate my mean in the while loop, it creates a mean for every matrix, and if I do it outside the while loop, it does it only for the last matrix.

For example while(reading file) ...

if(some conditions) continue

end

match= strcmp('Hello', out)//compares script with the word hello and outputs a logical array of 1s and 0s

M=mean(match,3)

end

match=[1 0 0 0 0 0] match=[0 0 1 0 0 0] match=[0 0 1 0 0 0]

I want the mean to be M=[1/3 0 2/3 0 0 0]

Do I need to use a for loop? Because it seems to be a much longer process to use a for loop to get the mean, but I do not see how else I could do this.

Answer 1

[ChristianW]

k = 0;
while (reading file) ...
    k = k+1;
    match(k,:) = strcmp('Hello', out)
end
M = mean(match)

[Samira]

I just tried your code, and it's giving me this error, and doesn't read past k=k+1;

??? Subscript indices must either be real positive integers or logicals.

But match is logical, so I don't understand why this error comes up.

[ChristianW]

Dont use variablename k for the counter/index if k is allready in use.

[Samira]

Thank you, I've kinda got it to work. It seem I have the wrong dimensions. My match is now a 2 by 5 logical array. When I'm testing this out, I'm using a script with 3 lines. But it doesn't read past the second line. This error comes : ??? Subscripted assignment dimension mismatch.

Obviously, I have the wrong dimensions, how do I fix this error?

[ChristianW]

Explain the diversity of your (read) file or the out-variable.

files = {{'Hello','bla','bla'}
         {'bla','Hello','Hello'}
         {'Hello','Hello','bla'}};
for k = 1:length(files)
    out = files{k};
    match(k,:) = strcmp('Hello', out);
end
M = mean(match)

[Samira]

This is my code so far

i=0

fid=fopen('test.m')

while~ feof(fid)

line=fgetl(fid);

if isempty(line)||strncmp(line, '%',1)||~ischar(line)

continue

end

fprintf(line);

out=regexp(line, ' ', 'split')

i=i+1

match(i,:)=strcmp('Hello', out)

end

M=mean(match)

match is a 2 by 5 logical cell and out is a 1 by 8 array. This one is just an example though, my actual code is reading a file with 100's of conversations between two people so it is a big file.

[ChristianW]

Your statements are full of contradictions.

Are the number for words per line allways the same? Because this ...

files = {{'Hello','bla','bla'}
         {'bla','Hello','Hello'}
         {'Hello','Hello','bla','Hello'}}; % input for my code above

... can not work the way you planned it. How do you want to deal with that input? Do you want a logical array like this:

match = [ 1  0  0  NaN
          0  1  1  NaN
          1  1  0  1 ]

or like this:

match = [ 1  0  0  0
          0  1  1  0
          1  1  0  1 ]

or something else?

[Samira]

My file has sentences of different length, so the number of words differ. I want it to be as the later.

files = {{'Hello','bla','bla'} {'bla','Hello','Hello'} {'Hello','Hello','bla','Hello'}}; I'm created a matrix for each sentence so my output is like this

 match=[1 0 0 0]
 match=[0 1 1 0]
 match=[1 1 0 1]

I want it this way so I can work out the average. Is the reason I'm having difficulty because of the difference in the number of words per matrix?

Would it work if I made all matrices the same length, in this case n=4?

[ChristianW]

No, that is not your output. Your output is like this:

match=[1 0 0]
match=[0 1 1]
match=[1 1 0 1]

And yes, thats very possible the reason for your difficulties. To fill the empty space with zeros, you could use:

ml = strcmp('Hello', out); % match for the line
match(i,1:length(ml)) = ml; %#ok<SAGROW>

[Samira]

Sorry, I meant this is what I want my output to be like. Thank you so much, it works perfectly now! I have been trying to get this code to work for some time, thanks for taking the time to help me out, much appreciated. :D

[ChristianW]

You're welcome.